In normal speech, adults fluidly utter 2 to 3 words per second selected from no less than 50,000 to 100,000 regularly-used words in the mental lexicon. Lexical selection is the process by which we access and fit an appropriate word to ongoing speech and is a core process for language production. Pathological disruption of lexical selection (i.e., anomia) is observed to some extent in all aphasic patients and can also be present in neurodegenerative diseases as well as normal aging. Despite the central importance of lexical selection to language and the immense personal and societal cost caused by its disruption, its neural basis and that involved in its recovery are poorly understood. Different regions of the prefrontal cortex (PFC) and of the left temporal cortex (LTC) have been associated with lexical selection but their precise role and the way they interact to enable lexica selection and compensation are still unknown. This lack of knowledge has both methodological and theoretical causes: (a) The brain imaging techniques generally used to study healthy speakers do not provide temporal information on the brain regions that are involved in lexical selection. Knowing when these different regions come online during lexical processing would confirm their different roles and would also suggest how they interact with each other to achieve lexical processing. (b) Language has typically been considered in isolation from other cognitive functions. A burgeoning body of evidence suggests that lexical selection is dependent on domain-general cognitive control mechanisms. If this is true, it would suggest that different brain regions play different linguistic and non-linguistic roles, a finding that would have significant ramifications for treatment and recovery of brain-injured patients. The goal of the proposed research is to test the hypotheses that: (1) Lexical selection is supported by top-down modulations from the left PFC onto the left LTC, similar to how non-linguistic action selection is controlled, and (2) The right PFC serves as a compensatory top-down mechanism for lexical selection when the left PFC is injured, similar to what has been shown outside of the language domain. To test these hypotheses, a combination of novel theoretical and methodological approaches will be used. First, the unique opportunity to record electrocorticography (ECoG) in neurosurgical patients will be used to determine which, when and how subregions of the PFC and the LTC are involved in lexical selection. Second, linguistic vs. domain-general contributions to lexical processing will be tested by directly comparing patients' performance on two tasks using the same stimuli, but differing only in the type of stimulus-response associations to be performed (linguistic vs. domain-general). Finally, these same tasks will be used with neurological patients recovering from stroke in either the left or right PFC in order to determine the effects of these injuries on lexical selection. Behavioral measurements as well as surface EEG signals will determine the relative roles of these regions in lexical selection, and whether the right PFC participates in compensating for the loss of top-down control when the left PFC is injured.